use whisper.transcribe

app.py

@@ -243,7 +243,7 @@ with tap_input:
         if wav_audio_data is not None:
             st.audio(wav_audio_data, format='audio/wav')
             bytes_to_wav(wav_audio_data, output_wav_file='output.wav', sample_rate=16000)
-            user_input = transcribe('output.wav')
+            user_input = WHISPER_MODEL.transcribe('output.wav', language=WHISPER_LANG)['text']
 
     if submitted or wav_audio_data is not None:
         st.session_state['user_input_content'] = user_input